Alibaba releases two open-sourced AI models to rival Meta’s Llama 2

Alibaba Group, the Chinese tech and e-commerce giant, declared the launch of two open-sourced artificial intelligence (AI) models from its cloud computing section on August 3rd, as per an announcement. Named Qwen-7B and Qwen-7B-Chat, each is endowed with 7 billion parameters, describing them as measly-sized variants of the Tongyi Qiawen. These two new models have been desgned to facilitate the adoption of AI for operations of small and medium-sized organizations.

Qwen-7B & Qwen-7B-Chat boasts various user-friendly features which could be appealing to businesses. For instance, their project codes, model weights, and supporting documentation could be readily accessed by commercial establishments, scholars and scientists all over the world freed of cost. Furthermore, these are the leading LLMs that highlighted by a Chinese tech corporation that have been solely put out in a “open-source” setting. In contrast, commercial entities having been more than 100 million dynamic monthly customers are obliged to acquire a license for use.

In yet another occurrence this July and while dealing with Microsoft,q Alibaba initiated the first training as well as authorizing resolution for the total Llama2 series located in China. Such revelation followed Microsoft’s releases of their innovated Llama 2, having been educated via 40% more aggregated public datums and process units counting twice as distinguished environment thenlst effects as it predecessors. It too was free sourced, characterized by the purest quantities of Llama 2 established with 70 billion parameters, above which businesses generating over 700 million dynamic user-based questionnaires were asked to acquire a classified license.

Robert Wilson author
Articles: 12200